Biologist

$70–90/hrrate as listed

micro1 recruits biologists to write and grade problems that test whether AI models can reason correctly about molecular biology, genetics, physiology, and experimental design. Observed pay for this role runs $70–90/hr, with work delivered asynchronously against a weekly task queue.

Summary, eligibility guidance and screening preparation are prepared by Sidequest from the platform's posting. Check the source listing before applying.

What the work involves

You write questions a language model should be able to answer if it genuinely understands biology — and then you grade what it produces. Tasks vary by project: authoring graduate-level prompts with defensible reference answers, comparing two model responses and writing a rationale for which is stronger, flagging fabricated citations or invented mechanisms, or annotating where a model's reasoning chain breaks down even when its final answer happens to be right. Some batches lean toward experimental interpretation (here is a figure, here is a claim — is the claim supported?), others toward factual precision in areas like pathway regulation, immunology, or population genetics.

The grading rubric is usually stricter than a journal reviewer's, because the goal is to separate plausible-sounding text from correct science. Reviewers who write "looks fine" get filtered out quickly; the value you add is in the specificity of your critique — naming the wrong enzyme, the misapplied control, the conflated timescale.

What the platform screens for

Verifiable graduate training in a life-science discipline (PhD, or MSc plus research output) and a subfield you can be pressed on
Ability to state a claim, cite the basis for it, and hold up under two or three rounds of follow-up questioning
Written reasoning that a second expert could audit — not just verdicts
Comfort saying "the literature is unsettled here" rather than inventing certainty

micro1's screening is AI-led: a structured interview with adaptive follow-ups, generally followed by a paid or unpaid sample task in your stated area.

Logistics

Fully remote and asynchronous. Contributors typically commit 10–20 hours per week, though volume fluctuates with project cycles and can pause between batches. Work is contractor-based, invoiced hourly or per completed task depending on the project. Expect an initial calibration period where your first submissions are reviewed closely before you're given higher-trust or higher-rate queues.

What they look for

Biology
Molecular Biology
Wet Labs
Crispr
Cell Biology

Before you apply

Would you pass this screen?

Drawn from what this listing actually requires. Answered in your browser — nothing is sent to us, and we do not store your answers.

Do you have a PhD in a life-science discipline, or an MSc with published research output (as first author or major contributor on a peer-reviewed publication)?
Graduate training in life sciences · required
Can you name a specific subfield or domain within your training where you have hands-on research experience (months of direct work, not reading) and could explain in detail if asked?
Subfield expertise · required
Are you comfortable writing detailed, technical feedback in English that other experts in your field could audit and understand?
English written communication · required
Can you commit to a minimum of 10 hours per week, understanding that volume fluctuates between projects and there may be gaps between batches?
Availability · preferred
Do you have reliable internet and a quiet workspace to work fully remotely and asynchronously on written evaluation tasks?
Remote work and asynchronous tasks · required

0 of 5 answered

Included with this listing

Interview prep — 48 questions

Matched to what this platform's screening actually measures. Answer each out loud in 60–90 seconds: claim → concrete example → implication.

What it tests: Whether you can reconstruct a real scientific narrative with specific methods, controls, and reasoning rather than reciting textbook answers. This tests whether your domain knowledge is rooted in hands-on research or secondhand.
A strong answer: Candidate names a specific organism, technique, or system they worked with; describes what hypothesis or phenomenon they were testing; explains what the key measurement or observation was; and acknowledges what the data actually showed, including limitations or unexpected results.
Where people lose it: Vague recollection ('I studied cancer cells') without naming the specific pathway, cell line, or readout you measured. Or skipping the control or the interpretation step — just stating 'we found X' without explaining why that answers the question.